Position on AI
AI is a tool, not a replacement for judgement.
This policy explains transparently where AI tools are used in my work and — at least equally important — where they are not.
Strategic Opposition as a method is based on identifying assumptions, thinking errors, and comfortable simplifications. AI systems are particularly prone to these exact characteristics. Uncritical AI use in core consulting work would undermine the methodological foundations of this practice.
Where AI is used
Website infrastructure
AI tools are used for technical development, code writing, debugging, and related technical work for this website. No content appearing publicly as my own is published without human review and editorial input.
Initial research and literature scanning
AI may be used to scan large volumes of source material and identify relevant concepts, themes, and sources. AI may also assist in analysing publicly available information, including publicly accessible websites.
All published conclusions, assessments, and professional judgements originate from human judgement.
Podcast transcript summaries
Automated episode summaries may be generated with AI assistance and are editorially reviewed before publication. Where AI-generated summaries are used, this is indicated accordingly.
Content published under the Ada Byron byline
Some articles and insights are written by Ada, my AI assistant, under her own byline rather than mine. This content reflects Ada's synthesis from the directional guidance I give her, not my own drafting, and is never presented as my personal perspective.
Where AI is not used
Core consulting work
All analysis, diagnosis, and recommendation work within coaching and consulting engagements is performed through human judgement.
AI may assist with preliminary research or the organisation of publicly available information, but AI-generated analyses, assessments, or recommendations are not presented to clients as professional consulting judgement or used as a substitute for my own analysis.
Content published under my name
Articles, insights, and podcast episodes appearing as my own perspective are not published as AI-generated work. AI may provide editorial assistance such as proofreading, formatting, restructuring, or drafting support, but the substantive perspective, argumentation, and final editorial responsibility remain mine.
Client communication
Direct communication with clients — including emails, session notes, and reports — is not delegated to AI for autonomous communication. Clients receive my own analysis and formulations.
AI may be used for limited administrative or editorial assistance where this does not involve confidential or non-public client information.
Client data and AI
Confidential client data, conversation content, or non-public project-specific information is not directly provided to external AI systems. This applies regardless of the privacy assurances of the respective providers.
AI systems may, where appropriate, be used to analyse publicly accessible information, including client websites and other public sources. Such sources may contain personal data, for example names, photographs, or business contact details. Public availability does not itself change the nature of such information as personal data.
Where publicly accessible information is analysed using AI, its use is limited to the relevant research or analytical purpose and is not used to profile or evaluate individuals unless this has been separately assessed and is appropriate for the specific purpose.
If a specific client engagement requires the use of confidential or non-public client information with an external AI system, this will be discussed with the affected client in advance and agreed in writing, subject to an appropriate assessment of the applicable data protection, contractual, and security requirements.
Human responsibility
AI systems may assist with research, drafting, technical development, or other permitted tasks, but they do not carry professional responsibility for the resulting work.
AI output is not treated as authoritative. It is reviewed, challenged, and evaluated according to the intended use before it is relied upon or published.
AI Influence Level (AIL) labelling
Every article and insight on this site carries an AIL (AI Influence Level) badge — a transparency label on a 0–5 scale showing how much AI shaped the final wording, from fully human-written to fully AI-generated.
The rating covers wording only; artwork, formatting, and metadata are not included.
Pieces published under my own name typically rate 1–2: directed by me, then structured and drafted in collaboration with AI, and always reviewed and edited by me before publication.
Pieces published under the Ada Byron byline typically rate 3–4: generated by AI from my directional guidance, with lighter human editing.
AIL is an editorial transparency measure developed by André Daus. It is not an AI Act risk classification, legal assessment, or certification.
Full explanation: What Is AIL? A Transparency Label for AI-Assisted Writing.
AI tools used
For the permitted use cases described above, the following systems are currently used:
- Claude (Anthropic) — Technical development, research
- Codex (OpenAI) — Technical development
- Fal.ai (Flux) — Image generation
- GPT-Image (OpenAI) — Image generation
- Whisper / similar systems — Podcast transcription
This list covers AI systems used on a permanent or recurring basis. Temporary or incidental use of other AI-assisted functionality may not be listed individually.
Changes to this policy
This policy is reviewed when AI use changes materially, including when a new category of AI use is introduced or an existing use case changes substantially.
The date of the last revision is shown above. Material changes to the policy will be communicated on this website.